Autopilot steering off course

[V-Dog]

I wanted to give an update and thank all those that provided guidance.

I found no errant metal objects near the fluxgate compass.... after I finally found the fluxgate compass. It became an advanced game of Where’s Waldo?

When we had steady cellphone service I was able to update the software, which was a few versions out of date. I ran the Wizard Setup program and it kept getting stuck on the “checking rudder alignment” section. A few phone calls to “Ray” and it looks like the rudder feedback unit may be the culprit. I unfortunately had to leave the boat (damn job!!!) and won’t be back until the end of the month, so the issue still hasn’t been resolve yet. I’ll let you know how that goes.

Now it’s time to work on my mono/catamaran, best anchor, guns on board post...

Thanks again for your help.
Andrew

[tweekes]

Had the same problem with my old ST5000+, rudder transducer went bad.

[Peeew]

I have a similar setup on my boat. And I will reset the cross track error immediately before engaging a waypoint mode. My flux gate was way off at one point when I bought the boat and I needed to physically rotate the flux gate in its bracket to get it close. And the autopilot has a initial startup calibration function that might occasionally need to be redone.
Good luck and welcome to the forum and the Chesapeake Bay.

[pendragon]

Thew rudder feedback unit is probably worn in the center position. The center position is where it operates most of it's working life.
It is just a wire wound pot meter with center=zero . Replace it with a new one,

[fcallard]

I had a problem with a 30° error and it turned out to be due to some steel-framed beach lounges that were store in a quarter berth neat the flux gate compass. The marine electronics tech that trouble shot my system pulled out everything—including cushions(!) near the flux gate compass and used a hand bearing compass to sweep the removed gear. That turned up the offending beach lounges. We moved them away and then swung the compass and all was well.

[dfelsent]

Oh yes and don’t forget about the cast iron sewing machine

[pa391as]

Originally Posted by Dsanduril

Well, the only other thing I have up my sleeve is track mode, in track mode the Ray AP (at least the SPX series) will turn exactly 30 degrees if the XTE is > 0.01 nm until it gets XTE down to 0.0. This is so bad that it makes track mode almost useless. Doesn't sound like you're using track mode, but when I see 30 degrees in reference to a Ray AP, well, might as well mention it. I've had a bit of conversation with Ray about both of these "features" and they are pretty unapologetic, while acknowledging that they do exist.

Before disconnecting the rudder sensor I'd probably run through the calibration as Bill suggests. It's pretty simple and may sort out the problems.

I had the same experience this past weekend. My setup uses OpenCPN running on a tablet to send autopilot commands via NMEA0183, through a ShipModul Miniplex, to the NMEA2000 network where the Raymarine AP plays. I noticed ther same thing you reported, which is that if I wasn't right on course, the AP would change course to most aggressively get XTE to 0.00. This was a bit disconcerting because a human would normally set a course with a bearing to the next waypoint, not veer port/sb to get immediately back on course. there should be a setting like "immediately on course" or "on course to next waypoint"

[Dsanduril]

Yes, it can be disconcerting Especially because it frequently results in oversteer and a doubly-disconcerting turn back the other direction. Downwind it can downright dangerous. I’d like to see a “maximum turn angle” setting that I could put to 5 or 10 degrees. I’ve raised that point with Ray, they blew me off.

The answer may be pypilot but I haven’t got there yet.

[CassidyNZ]

Originally Posted by pa391as

My setup uses OpenCPN running on a tablet

Which tablet?

[boatbod]

The other possibility for the OP's problem is that the steering raw wiring has been connected with reversed polarity. In other words when the AP wants to turn the rudder to port it goes to starboard instead. When you first press Auto, the pilot may make attempt a small course correction but it sees the rudder go the wrong way so it tries harder to make the correction and the end result is a sharp turn. Suggest going through the dockside commissioning to make sure the rudder sensor and ram work properly.


Next up you can check your fluxgate/heading sensor but motoring around in a slow circle and confirming that the reported heading runs through a full 360 degrees and doesn't get "stuck" in a particular quadrant.